A Survey on Recursive Self-Improvement in Healthcare AI Agents
Published in alphaXiv preprint, 2026
Healthcare agents are increasingly built to refine their own policies rather than to run a fixed pipeline. This survey separates what an agent changes about itself, whether persistent memory, reusable skills, training material or the procedures that generate future improvements, from performance benefit and clinical validation, and maps the evidence across 85 studies.
alphaXiv · Project site · GitHub · Preprints.org v1 (first released as Self-Evolving Agent Engineering for Healthcare: Methodologies and Applications)
Recommended citation: Hou, D., Wu, Z., Zeng, Y., Jiang, L., Lin, F., & Yamada, K. D. (2026). A Survey on Recursive Self-Improvement in Healthcare AI Agents. alphaXiv.
Read Paper